ECG based assessment of circadian variation in AV-nodal conduction during AF—Influence of rate control drugs

Figures

• 

  FIGURE 1. A flowchart of the overall framework for estimating AV node conduction properties on an individual and a population level.

• 

  FIGURE 2. A schematic representation of the network model where the yellow node represents the coupling node, the red nodes the SP, the green nodes the FP, and arrows the direction for impulse conduction. For readability, only a subset of the 21 nodes is shown.

• 

  FIGURE 3. Mean (colored lines) and standard deviation (colored areas) of the error ϵ for 100 segments for the proposed genetic algorithm (blue) and a standard genetic algorithm (red) together with the Poincare difference ΔP(i) (black line), defined in Eq. 5, for data from one patient at baseline during 3 hours. The standard deviation and mean are based on ten runs of the algorithms. Note that ΔP(i) is scaled with 15 for readability.

• 

  FIGURE 4. The Poincaré plot with associated histogram and RR interval series of data (blue) and model output (orange) for the θ^(i) corresponding to the median ϵ (left) and to the θ^(i) which ϵ is higher than 95% of all ϵ (right).

• 

  FIGURE 5. Estimated model parameters θ^(i) (black dots), with corresponding uncertainty estimates PU (red areas), along with the fitted cosine (blue lines) used for the circadian variation, for one patient during baseline. In each panel, the RMSE is reported as a measure of goodness of fit between θ^(i) and the fitted cosine. Left column shows the parameters relating to the minimum conduction delay or refractory period, the middle column the parameters relating to the maximum prolongation, and the right column to the time constants. For further explanation of the model parameters, see Table 1.

• 

  FIGURE 6. Mean (circles) ± one standard deviation (bars) of the parameter uncertainty estimates u(p) over all recordings and all patients, normalized with the parameter ranges r(p). Note that the model parameters RminSP and ΔR SP have a lower uncertainty, indicating a higher impact on the resulting model outcome.

• 

  FIGURE 7. The fixed effects with corresponding 95% confidence intervals for the cosinor mean m (top) and cosinor amplitude a (bottom) for each model parameter (cf. Table 1) and drug. Confidence intervals not overlapping zero indicate significant difference from baseline (p <0.05).